Method Pack · M9 · Interactive

Building an MEL System from Scratch

A 3-hour interactive Practice Pack. You have joined a programme that has been running for two years with no formal monitoring. Walk through six months of setup and walk out with a 90-day MEL plan you can actually execute.

4 modules~3 hoursInteractiveIndia-context

Your progress
0% complete
Your Capstone

90-Day MEL Setup Plan

Walk in with an organisation that has no formal MEL. Walk out with a phased 90-day plan covering audit findings, theory of change, indicator framework, data collection tools, and learning-loop design. Built automatically from your module answers.

Module 1 . ~25 min

Month 1: Audit what exists

Every organisation already monitors something, even if nobody calls it MEL. The first month is about discovering what already exists, what is being collected informally, and where the gaps actually are. Resist the temptation to design a new system before understanding the old one.

The five places data already lives

The audit conversation

Sit with three people individually: (a) the programme director, (b) a senior field coordinator, and (c) a data-entry person if one exists. Ask each of them:

  1. What information do you collect regularly? (probe: registers, forms, apps)
  2. What do you do with it after collecting? (probe: who reads it, what decisions does it feed?)
  3. What do you wish you knew but currently do not?
  4. What reporting burden do you find pointless?
Illustrative example (composite case, hypothetical figures)

An Odisha livelihoods NGO (a composite case, not a named organisation) ran an NRLM-linked livelihoods programme that had been running for three years with no MEL officer. The audit found: (1) SHG meeting registers with attendance and savings data in paper form across 120 villages. (2) Monthly field reports submitted by area coordinators as Word documents with inconsistent formats. (3) NRLM MIS data entry happening at the block level with a two-month lag. (4) Zero outcome-level data.

Key insight: The SHG registers were the richest dataset -- just never aggregated. The new MEL system started by digitising those, not by creating a new survey instrument.

Your MEL Audit Sheet

Fill these for your organisation. Your answers save automatically and flow into the final capstone.

e.g., "Odisha livelihoods NGO -- NRLM-linked SHG livelihoods programme, 120 villages, Odisha, 3 years running"
Registers, MIS, reports, WhatsApp groups, Excel sheets, government reporting
e.g., NRLM MIS, UDISE+, HMIS, PFMS, Awaas+, PM-KISAN
Saved
Self-check
You join a nutrition programme in Jharkhand that has been running for two years. The team says "we have no data." What is your most productive first step?
Design a new baseline survey
Hire a data entry operator
Ask to see funder reports, field registers, and the ICDS/Poshan Tracker data they already submit
Set up a KoboToolbox account
Correct. "We have no data" almost never means zero data exists. It means nobody has aggregated or used it. The audit conversation reveals what already exists before you add new collection burden.
Module 2 . ~30 min

Month 2: Theory of Change + Indicator Framework

The theory of change is not a wall poster. It is the backbone of your indicator framework. Every indicator you select must connect to a specific causal link in the ToC. If an indicator does not test a causal assumption, it is a vanity metric.

Building the ToC in practice

Run a half-day workshop with the programme team. Use this sequence:

  1. Start from the right -- what is the long-term change this programme seeks? (e.g., "Women in 120 villages have sustained increase in household income and economic agency.")
  2. Work backwards -- what intermediate outcomes must be true for that to happen? (e.g., "SHG members access formal credit," "Micro-enterprises survive beyond 12 months.")
  3. Identify outputs -- what does the programme directly deliver? (e.g., "SHG formation and training," "Enterprise development support.")
  4. Map assumptions -- what must be true between each link? (e.g., "Banks will lend to SHGs," "Market demand exists for the products.")

From ToC to indicators: the three-tier framework

TierWhat it measuresFrequencySource
Tier 1: Activity trackingInputs and outputs (training conducted, SHGs formed, loans disbursed)MonthlyMIS, field reports
Tier 2: Outcome monitoringIntermediate changes (repayment rates, enterprise survival, income change)QuarterlyStructured follow-up, government data
Tier 3: Impact signalsLong-term change (economic agency, food security, asset ownership)Annual or evaluation-onlySurveys, qualitative deep-dives

The "20 indicator" ceiling

Programmes with several funders tend to accumulate long indicator lists, because each funder adds its own. A workable MEL system monitors a short core set routinely. The rest stay in the logframe but are not part of the routine data collection. If your field team cannot explain what each indicator means and why it matters, cut it.

Government scheme alignment

If your programme operates under NRLM, PMAY-G, Samagra Shiksha, or NHM, the government scheme already has a results framework. Do not duplicate it. Align your Tier 1 indicators to the scheme's reporting requirements and add only the Tier 2-3 indicators that the scheme framework misses.

Your ToC and Indicator Framework

Build the skeleton. These flow into your capstone.

What must be true for your ToC to hold? List 2-3 testable assumptions.
Saved
Self-check
Your indicator framework has 52 indicators. The field team can reliably collect 18 per month. What is the best response?
Hire more data collectors
Prioritise 15-18 for routine monitoring, move the rest to periodic or evaluation-only collection
Automate all 52 using a mobile app
Report all 52 but only verify a sample
Correct. Data quality collapses when collection burden exceeds capacity. The solution is tiering, not adding staff or technology. Technology helps with the 18 you keep, not with justifying 52 nobody reads.